The goal is to make this library take advantage of the native APIs like URLSession and Android's Ktor engine for better performance and reliability.
If you would like to contribute, please join our Discord and ask questions in the #oss channel or create a pull request.
Requires iOS 15+
and Android minSdkVersion = 24
.
yarn add react-native-openai
import OpenAI from 'react-native-openai';
const openAI = new OpenAI({
apiKey: 'YOUR_API_KEY',
organization: 'YOUR_ORGANIZATION',
});
// Alternatively (recommended), you can use your own backend and not hardcode an API key in your app i.e. https://my-custom-domain.com/v1/chat/completions (follow the same API as OpenAI until pathPrefix is fixed on Android)
const openAI = new OpenAI({
host: 'my-custom-host.com',
});
const [result, setResult] = React.useState('');
// Listen for messages
openAI.chat.addListener('onChatMessageReceived', (payload) => {
setResult((message) => {
const newMessage = payload.choices[0]?.delta.content;
if (newMessage) {
return message + newMessage;
}
return message;
});
});
// Send a message
openAI.chat.stream({
messages: [
{
role: 'user',
content: 'How do I star a repo?',
},
],
model: 'gpt-3.5-turbo',
});
// Alternatively, you can use the create method if `stream` is false
await openAI.chat.create(...)
const result = await openAI.image.create({
prompt: 'An awesome Candle logo',
n: 3,
size: '512x512',
});
setImages(result.data.map((image) => image.url));
Thank you to Dylan Shine & Mouaad Aallam for making openai-kit and openai-kotlin which this library is based on.
MIT